Re: LSB.fhs failure digest for Debian Sample Implementation
On Aug 20, 2:29pm in "Re: LSB.fhs failure ", Matt Taggart wrote:
> I guess my point is that FIP results are time consuming since you have to go
> through and look at them(or add the manual declaration mentioned above). It
> would be nice it the test knew how to deal with these common problems.
> Eventually the test suite should be able run from cron and only failures
> mailed off for people to look at. Having to do these be hand is a PITA.
If the LSB had added some way to automatically determine the FHS options
supported by the implementation then we could have done this. As it
is the FHS tends to have statements saying "if the subsystem is supported".
As such this way of testing is one of the differences between producing
a certification test suite and a regression test suite. For regressions
you'd store a journal of expected results and use the comparative
reporting facility in vrpt just to flag the changes between that which
was expected for the platform.
We could add a large number of questions up front to the configuration
but wanted to keep it simple for setup. We do at least need to
add some more documentation on these type of issues in the READMEs.
regards
Andrew
Reply to: